Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Mon, 12 Jul 2021 19:10:14 +0800
From:   Leo Yan <leo.yan@linaro.org>
To:     Suzuki K Poulose <suzuki.poulose@arm.com>
Cc:     Mathieu Poirier <mathieu.poirier@linaro.org>,
        Mike Leach <mike.leach@linaro.org>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        coresight@lists.linaro.org, linux-arm-kernel@lists.infradead.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/2] coresight: tmc-etr: Correct memory sync ranges in SG
 mode
Message-ID: <20210712111014.GC704210@leoy-ThinkPad-X240s>
References: <20210710070206.462875-1-leo.yan@linaro.org>
 <3038eb59-155c-5c3d-18e3-e45e56b52523@arm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <3038eb59-155c-5c3d-18e3-e45e56b52523@arm.com>
Precedence: bulk

On Mon, Jul 12, 2021 at 11:25:54AM +0100, Suzuki Kuruppassery Poulose wrote:
> Hi Leo,
> 
> On 10/07/2021 08:02, Leo Yan wrote:
> > Current code syncs the buffer range is [offset, offset+len), it doesn't
> > consider the case when the trace data is wrapped around, in this case
> > 'offset+len' is bigger than 'etr_buf->size'.  Thus it syncs buffer out
> > of the memory buffer, and it also misses to sync buffer from the start
> > of the memory.
> > 
> 
> I doubt this claim is valid. We do the sync properly, taking the page
> corresponding to the "offset" wrapping it around in "page" index.
> 
> Here is the code :
> 
> 
> 
> void tmc_sg_table_sync_data_range(struct tmc_sg_table *table,
>                                   u64 offset, u64 size)
> {
>         int i, index, start;
>         int npages = DIV_ROUND_UP(size, PAGE_SIZE);
>         struct device *real_dev = table->dev->parent;
>         struct tmc_pages *data = &table->data_pages;
> 
>         start = offset >> PAGE_SHIFT;
>         for (i = start; i < (start + npages); i++) {
>                 index = i % data->nr_pages;
>                 dma_sync_single_for_cpu(real_dev, data->daddrs[index],
>                                         PAGE_SIZE, DMA_FROM_DEVICE);
>         }
> }
> 
> 
> See that the npages accounts for the "size" requested and we wrap the
> "index" by the total number of pages in the buffer and pick the right
> page.
> 
> So, I think this fix is not needed.

Ouch, you are right :)  Let's drop these two patches.

Thanks,
Leo